Purpose of the Study: The primary purpose of this study is to enhance accuracy, sensitivity, precision, specificity and error rate for an Advanced Microwave Scanning Radiometer - Earth Observing System (AMSR-E) images. Methodology: K-Nearest Neighbor classification (KNN) method proposed and implemented in MATLAB for AMSR-E sea ice images to achieve high accuracy, sensitivity, precision, specificity and acceptable error rate. Main Findings: From the findings, the proposed method increases accuracy, sensitivity, precision, specificity and error rate by 12.1%, 79.9%, 10.0%, 10.0%, and 12.1%, respectively compared to non-linear SVM; 17.6%, 74.8%, 10.8%, 11.1%, and 17.6%, respectively as compared with decision tree method. Applications: The study results help develop concepts or theories for the earth observation system at Polar Regions and critical notes for global warming situations. Novelty/Originality: K-Nearest Neighbor classification (KNN) method is newly introduced on polar region images.